Replications of empirical studies can be a very valuable source of learning and progress in scientific research. Replications should be the basis or starting point of any research project that builds on other people’s work. My posts on replications below are meant to support the students and the courses I teach at WU Vienna, and to contribute to critical (academic and public) discussions in a much broader context. This second goal agrees with the intentions of the Replication Network: “Why replications? They enable researchers to (i) determine the fragility or robustness of previous research findings, and (ii) identify why studies reach different conclusions on the same subject.” Gary King has highlighted the need and value of replications in two papers (1995, 2006). He states “… scientists needs access to the body of data necessary to replicate existing studies to understand, evaluate, and especially build on this work. […] The only way to understand and evaluate an empirical analysis fully is to know the exact process by which the data were generated and the analysis produced.”
Replications have typically not been highly appreciated in academic research. This view has changed in the past decade or so, and replications have received considerably more attention. This change can be viewed as a response to the so-called ‘reproducibility crisis‘, and the increasing awareness about (and/or the negative effects of) p-hacking or data dredging. Academic journals have started to respond to these challenges by requesting authors to provide data and source code of their analyses, and have changed their submission guidelines. For example, guidelines of the Econometric Society impose restrictions on the use of asterisks to highlight significant estimates, and suggest using confidence intervals. The Strategic Management Journal “no longer accepts papers for publication that report or refer to cutoff levels of statistical significance (p-values)” (author guidelines). The Critical Finance Review “insists that authors make their programs and/or data available“, and its editor “believes that the hallmark of a science is not replicability but critical replication.”
In my posts I attempt to write in way that is both accessible to a broader audience (without methodological skills or interests in those skills), and for those who would like to acquire (additional) methodological skills. I will typically avoid formal derivations or explanations here, and refer to my lecture notes available from my WU website for further reading.
My replications and extensions:
Summaries of replications:
Alois Geyer 